546 
CL VII. GRAMINEJE (Stapf). 
[ Bmchiaria . 
113. Dembelos, Pappi, 6061. Ocule Cusai, 5250 ft., Pappi, 1809, 3995. I)am- 
veita ; Hotha Islands, Terracciano, 810, 811. Um Namus Island, Terrac- 
ciano, 430. Abyssinia: Tigre ; Hamedo plains, 4600 ft., Schimper, 1074! 
1088 ! Samen ; Tacazze Valley, Schimper, 1513 ! 1653 ! and without precise 
rocaTity, Schimper, 269 ! South Abyssinia ; Hamara on the Gonale River, 
Riva, 1195! and without precise locality, Drake-Brockman, 65! British 
Somaliland: without precise locality, Drake-Broclcman, 411! Italian Somali- 
land; Jub Valley, Erlanger. Socotra, Balfour, 23! 51! 67! 127! South 
of Tamariu, Schweinfurth, 429 ! 
Lower Guinea. Angola : Loanda ; gregarious, appearing after the rains, 
Gossweiler, 1655 ! 1667 ! in poor pasture, Welwitsch, 7467 ! 7362 partly ! 
Golungo Alto (?), Welwitsch, 7217! 7239c! Rondo, Mechow, 2! 2367! 
Mossamedes ; common undergrowth in open bush at km. 107 of the Mossa- 
medes railway, Pearson, 2901 ! Ramaraland, Een ! 
Mozambique Distr. Zanzibar, Ilildebrandt, 1086 ! Belligny ! Last ! German 
East Africa: Tanga Station, Holst, 2011 ! Ukambani; Kibwesi, in grassland 
and laterite, Scheffler, 523 partly ! Nyasaland : Milauri, among taller grasses 
not far from the Shire, Scott ! Portuguese East Africa : Senna, Carvalho ! 
Tette, Kirk ! Muchukwana on the lower Buzi, Swynnerlon, 997 ! Rhodesia : 
Victoria Falls, Rogers, 5720 ! 6006 ! 
Also in Yemen, N.-E. Transvaal (Shilouvane, Junod, 123 !) and Madagascar. 
Panicum nidulans, Mez, described from specimens collected by Schwein- 
furth on the shore at Mirza Elei, south of Elei Island (Nubia), and others 
collected by Ehrenberg (proba bly in Eritrea), is apparently merely a very hairy 
much stunted and contracted desert condition of B. regularis. Schweinfurth’s 
specimens much resemble in general habit certain specimens of the pubescent 
form of B. ramosa from N.-W. India, but as B. regularis seems to be common 
in Nubia and Eritrea whilst typical B. ramosa, appears to be absent there, we 
may be justified in connecting P. nidulans with the former; some of the 
primary pedicels are actually more elongated than one would expect from B. 
ramosa. A similar contracted dwarf state is represented by Broun’s specimen 
from Dakhlia Shellel, Sennar, and other approaches towards it are found 
among the Socotran specimens, of which Balfour says, ‘‘A very variable 
species, sometimes dwarf, from other localities a large stout plant.” The 
taller Socotran specimens are undoubtedly B. regularis, and are clearly 
connected with the dwarf states. A similar variation, evidently due to con- 
ditions of nutrition, is exhibited by Rogers’ specimens from the Victoria Falls 
(no. 5720, tall; no. 6006, stunted). No specimens referable to B. regularis . 
have been seen in Indian collections ; but as far as it goes Trinius’ figure C 
in Sp. Gram. Ic. t. 176 (P. Peiiveri ) seems to represent it. 
Chiovenda (l.c. viii, 32) based a ‘ ‘ var. robustissimum ” of Panicum Petiveri 
on Pappi 3418 from Mount Urug. He says it has the habit of P. maximum, 
with very robust, pubescent culms, a large panicle with the lowest branches 
in whorls of 5-7 and densely pubescent spike lets. A specimen of Pappi 3418 
in the Zurich Herbarium did not strike me as different from B. regularis as 
understood here. Chiovenda also quotes typical P. Petiveri from the same 
locality and collected on the same date (no. 3411). 
35. B. ovalis, Stapf. Annual, fascicled, up to over 1 ft. liigli. 
Culms erect or genicu lately ascending, about 4-noded, branched 
below (sometimes copiously), pubescent below the nodes, terete or 
compressed below. Leaf-sheaths thin, rather loose, striate, softly 
pubescent all over ; ligule a dense fringe of short hairs ; blades linear 
from an equally wide base, long-tapering to a fine point, 3-7 in. by 
2-3 lin., flat, flaccid, pale green, very finely and softly pubescent 
all over, margins finely cartilaginous, scabrid, not crisped, midrib 
